Proactive analysis of construction equipment operators' hazard perception error based on cognitive modeling and a dynamic Bayesian network

Construction equipment-related accidents are unarguably one of the most frequent types of construction accidents. Construction equipment operators’ hazard perception error (HPE) has been recognized as one of the primary reasons for these accidents. Operators’ hazard perception involves a series of cognitive functions that will change as the construction process evolves. Although the importance of hazard perception to construction equipment operating safety is widely recognized, the analysis and interpretation of its cognitive processes, potential cognitive failure modes, underlying causes, and dynamic characteristics involved has not been fully addressed. Furthermore, there is still a lack of an effective method to quantitatively assess HPE evolution and changes in corresponding cognitive states over time. This study combines a cognitive model and dynamic Bayesian network (DBN) modeling to provide a qualitative and quantitative proactive analysis of operators’ HPE. Considering the lack of prior knowledge of operators’ HPE in the construction industry, computational models of several key cognitive functions and multiple information sources were integrated to determine the conditional probability distributions of the DBN nodes. The method's feasibility was validated with a case study. Researchers and practitioners may customize the model to quantify the occurrence tendency of operators’ HPE under a specific construction condition to assist in proposing countermeasures to reduce and mitigate HPE.